AI designed drug alters biological aging markers in clinical trial

Inside a lecture hall at Sorbonne University in Paris, slide presentations from biotechnology executives usually focus on narrow clinical endpoints like forced vital capacity or lesion reduction. But when Insilico Medicine Chief Executive Officer Alex Zhavoronkov presented data at the Nature conference on redefining healthcare, the focus shifted to systemic biomarkers. The presentation, which followed a peer reviewed paper in Nature Biotechnology, detailed how a drug candidate designed by artificial intelligence altered the biological age profiles of patients during a clinical trial.

The drug, rentosertib, was developed to treat idiopathic pulmonary fibrosis, a progressive lung disease. While the primary objective of the Phase IIa trial was to evaluate safety and tolerability in treating lung tissue, researchers also analyzed blood protein data from the forty-two participants. Using computational models of aging, the research team found that patients treated with the molecule showed a reduction in their predicted biological age compared to those who received a placebo.

To measure systemic physiological changes, the researchers analyzed proteomic aging clocks. These models measure the concentrations of specific proteins in blood plasma to estimate an individual's biological age, which reflects cellular and tissue health rather than chronological years.

Instead of relying on a single proprietary model, the investigators analyzed patient blood samples using six distinct proteomic aging clocks. These computational models were developed independently by separate scientific institutions, including researchers associated with Harvard University, Oxford University, and Peking University, alongside Insilico's own algorithms. Each clock was trained on different population datasets and prioritized different protein pathways.

Despite these design differences, the six models produced consistent directional results. Patients who received rentosertib exhibited proteomic profiles that translated to lower biological ages than patients in the control group. According to the published study, the observed reduction in biological age averaged three to four years across the models, with one clock estimating a reduction of up to six years. The scientific significance of these findings rests on this multi model consensus, which suggests a systemic biological response rather than an artifact of one specific algorithm.

Algorithmic target identification and molecular design

The development of rentosertib represents an application of generative artificial intelligence to both target discovery and molecular generation. The process began with the identification of TNIK, a protein kinase associated with several molecular processes that drive cellular decline, often referred to as the hallmarks of aging.

Once TNIK was identified as a viable therapeutic target for both lung fibrosis and systemic cellular decline, Insilico used generative chemistry models to design a novel molecule engineered to bind specifically to this protein. The resulting compound was rentosertib. By inhibiting TNIK, the drug aims to resolve localized fibrotic tissue while simultaneously modulating broader cellular pathways. The alterations observed in the blood protein profiles of the trial participants suggest that targeting this single kinase may exert downstream effects throughout the body.

Commercial landscape and strategic pipelines

The integration of systemic longevity metrics into traditional clinical drug development points to a structural shift in how biotechnology companies design and license therapeutics. Historically, drug development has focused on a single disease indication, ignoring broader systemic decline. Insilico's strategy suggests a dual path where a candidate is advanced for a specific regulatory indication, such as lung fibrosis, while gathering data on systemic aging biomarkers.

This dual approach carries commercial value for developers. Biopharma companies that can demonstrate systemic health benefits alongside primary clinical efficacy may establish stronger differentiation in competitive therapeutic markets. Insilico has built its pipeline around this methodology, securing licensing agreements with global pharmaceutical firms including Eli Lilly, Servier, and Takeda to co develop and commercialize AI discovered assets.

For the broader industry, this model could reshape clinical trial design. If early stage trials routinely monitor systemic aging biomarkers, drug developers can identify secondary indications far earlier in the clinical lifecycle, reducing the capital risk associated with long term development pipelines.

Despite the consistent signals across the proteomic clocks, the researchers noted several limitations in the current dataset. Because the trial was conducted in patients with active idiopathic pulmonary fibrosis, the study cannot definitively separate systemic biological rejuvenation from the localized healing of the lungs. A patient experiencing improved respiratory function and reduced inflammation will naturally show improved systemic health markers, which proteomic clocks may interpret as a younger biological age.

To confirm a direct, systemic anti aging effect, researchers must eventually evaluate the molecule in healthy volunteers. A clinical trial in a healthy population would measure biological age changes without the confounding variable of active tissue disease.

For now, the clinical development of rentosertib is focused on its primary therapeutic indication. A Phase III trial is currently underway in China to evaluate the efficacy of the drug in a larger patient cohort with idiopathic pulmonary fibrosis, which will provide the definitive safety and efficacy data required for regulatory approval.